Short tail with skin lesion phenotype occurs in transgenic mice with keratin-14 promoter-directed expression of mutant CXCR2

Abstract

CXCR2 plays an important role during cutaneous wound healing. Transgenic mice were generated using the keratin-14 promoter/enhancer to direct expression of wild-type human CXCR2 (K14hCXCR2 WT) or mutant CXCR2, in which the carboxyl-terminal domain (CTD) was truncated at Ser 331 and the dileucine AP-2 binding motif was mutated to alanine (K14hCXCR2 331T/LL/AA/IL/AA). Our results indicate that K14hCXCR2WT transgenic mice exhibited a normal phenotype, while K14hCXCR2 331T/LL/AA/IL/AA transgenic mice were born with tails of normal length, but three to eight days after birth their tails degenerated, leaving only a short tail stub. The tissue degeneration in the tail started between caudal somites with degeneration of bone and connective tissue distal to the constriction, which was replaced with stromal tissue heavily infiltrated with inflammatory cells. The tail lesion site revealed coagulation in enlarged vessels and marked edema that eventually led to loss of the distal tail. Moreover, 66% of the mice exhibited focal skin blemishes and inflammation that exhibited an increase in the number of sebaceous glands and blood vessels, enlargement of the hair follicles due to increased number of keratinocytes, reduction in the connective tissue content, and a thickening of the epidermis. Furthermore, immunohistochemical staining of the epidermis from tail tissue in the transgenic mice indicated a loss of the cell adhesion markers E-cadherin and desmoplakin. These data suggest that keratinocyte expression of a CTD mutant of CXCR2 has effects on homeostasis of the connective tissue in the tail, as well as the maintenance of the epidermis and its appendages.

INTRODUCTION

Wound healing is a complicated but well-organized procedure that involves several overlapping phases: inflammation, cell migration, cell proliferation, neovascularization, extracellular matrix production, and tissue remodeling. In an acute wound, healing is regulated by many factors. Chemokines are among these regulatory factors involved in the wound-healing process [1,2,3,4,5,6]. Chemokines, small chemotactic proteins, are divided into four families: C, CC, CXC, and CX3C, based upon differences in conserved cysteine amino acid residues: the CXC chemokine subfamily is further divided into ELR⁺ CXC chemokines and ELR⁻ CXC chemokines, due to the presence or absence of the ELR motif (glutamic acid-leucine-arginine sequence) [1, 7,8,9]. The ELR⁺ chemokines are usually angiogenic, while the ELR⁻ chemokines are usually angiostatic.

The chemokine receptor, CXCR2, is a seven-transmembrane G protein-coupled receptor that binds the ELR+CXC angiogenic chemokines [9, 10]. CXCR2 function is important for keratinocyte migration and for excisional wound healing. In our previous studies, we demonstrated that CXCR2 knockout mice exhibit delayed wound healing, impaired neutrophil recruitment, and reduced angiogenesis in the wound bed. We also observed that the CXCR2 small molecule inhibitor (SB 225002, Smith-Kline Beecham, now Glaxo/Smith-Kline) markedly impaired wound healing in wild-type mice. Thus, it is clear that loss of function of CXCR2 results in impaired cutaneous wound healing [4].

After ligand stimulation, CXCR2 undergoes rapid coupling and uncoupling with the trimeric G protein and activation of a host of downstream intracellular signals. Trafficking patterns and molecular signaling cascades of CXCR2 are regulated spatially and temporally. Ligand binding leads to receptor internalization and trafficking through the endosomal compartments where the receptor is either moved to the lysosome for degradation or through the recycling endosomal compartments and back to plasma membrane [11]. Several important intracellular adaptor proteins, such as AP2, dynamin, and β-arrestin 1 and 2, are needed for CXCR2 internalization and its downstream signaling [12, 13]. These adaptor proteins bind specifically to functional motifs of the carboxyl-terminal domain of CXCR2. We have previously demonstrated that truncation of the carboxyl-terminal domain of CXCR2 at S331 (CXCR2 331T) produces a receptor, which is much slower in ligand-mediated desensitization and is poorly internalized in some cell types [14,15,16]. Further mutation of this truncated receptor in the LLKIL motif impairs the association of adaptor proteins AP2 and HIP with CXCR2 and impedes receptor internalization in HL-60 cells, an event that also impairs ligand-mediated chemotaxis as monitored in assays using a modified Boyden chamber assay or a microfluidic device [14]. However, the role of receptor internalization and adaptor binding in chemotaxis varies considerably among chemokine receptors and cell types (15).

In this study, we aimed to characterize the role of the carboxyl-terminal domain of CXCR2 in keratinocyte homeostasis and wound repair in the epidermis. We examined the effect of keratinocyte expression of human CXCR2 truncated in the carboxyl-terminal domain (CTD) and mutated in LL/IL motif. CTD sequences classically associated with receptor desensitization, internalization, and adaptor binding. We examined the phenotype of two founders of each genotype and also examined cutaneous wound healing responses in nontransgenic mice, transgenic mice expressing the wild-type form of human CXCR2 (hCXCR2 WT), and human CXCR2 truncated at Ser 331 and mutated in the Leu-Leu-Lys-Ile-Leu motif (hCXCR2 331T/LL/AA/IL/AA) (Fig. 1a). CXCR2 transgene expression was under the transcriptional regulation of the keratin-14 promoter/enhancer.

Fig. 1.

Construct of hCXCR2 and its mutant form. (a) At the carboxyl terminal of construct of hCXCR2, the arrows show the positions of the mutations and truncation. (b) The sequences of the carboxyl-terminal of hCXCR2 WT and the CXCR2 331T/LL/AA/IL/AA mutant. For the K14hCXCR2 331T/LL/AA/IL/AA mutant, the carboxyl-terminal truncation of hCXCR2 was produced by introducing a stop codon at Ser 331 and the Leu-Leu and Ile-Leu amino acids from the LLKIL motif were mutated to Ala.

The identities of transgenic founders expressing the wild-type or the mutant form of human CXCR2 were verified by polymerase chain reaction (PCR), Western blot, and immunostaining of tissues. Our studies show for the first time that elevated keratinocyte expression of a form of CXCR2 mutated in the AP-2/HIP binding motif produces a phenotype that indicates abnormalities in maintenance of the integrity of the epidermis and tails, which may be associated with the loss of E-cadherin and desmoplakin expression, further emphasizing the importance of CXCR2 in processes involved in epidermal homeostasis.

MATERIALS AND METHODS

Construction of hCXCR2 wild-type and mutant mice

The cDNAs encoding hCXCR2 WT or hCXCR2 331T LL/AA/IL/AA [15] were inserted into the cluster of restriction sites between the K14 promoter region and β-intron gene of the pGEM^r-3Z vector (the kind gift of Elaine Fuchs, Rockefeller University). The constructs generated were K14hCXCR2 WT and K14hCXCR2 331T/LL/AA/IL/AA. For the K14hCXCR2 331T/LL/AA/IL/AA mutant, the carboxyl-terminal truncation of hCXCR2 was produced by introducing a stop codon at Ser 331; then the Leu-Leu and Ile-Leu amino acids from the LLKIL motif in the carboxyl domain of CXCR2 were mutated to Ala (Fig. 1b). PshBI and PvuI restriction enzymes were used to release the fragments of the transgene DNA. The gel-purified transgene DNA from these constructs was introduced into the pronuclei of fertilized eggs from C57BL/6 mice by microinjection using standard protocols in our Vanderbilt-Ingram Transgenic Core Facility. Potential C57BL/6 founder lines were identified by genotyping the progeny from the pseudo-pregnant females transplanted with the microinjected pronuclei. The status of founders was evaluated further by genotyping the offspring of potential founder progeny.

Genotyping by polymerase chain reaction

To perform genotyping, 0.5 cm of the tip of mouse tail was removed from anesthetized 4-wk-old mice and digested overnight in 500 μl of lysis buffer (500 mM Tris pH 8.0, 100 mM EDTA, 0.5% SDS) containing 25 μl 10 mg/ml Proteinase K at 60°C. Two phenol/chloroform/iso amyl alcohol (25:24:1) extractions were performed, and the DNA was precipitated in 0.3 M sodium acetate in 100% ethanol. The precipitated DNA was collected with a glass hook, rinsed two times in 70% ethanol and one time in 100% ethanol. After the DNA was air dried for 10 min at room temperature, DNA was dissolved in distilled H₂O (dH₂O). The sequences of the two primers used for genotyping are 5′ primer: 5′-AATTCTGGCTGGCGTGGAAAT-3′ and 3′ primer: 5′-GGGGCAGGATCCGTAACGCA-3′. For each PCR reaction, 1× PCR buffer (Sigma-Aldrich, St. Louis, MO, USA), 200×4 μM dNTPs, 3 mM MgCl₂, 0.2 μM of each primer, 2.5 units of Tag polymerase (Sigma-Aldrich) and 1 μg of DNA were used. The setting of the PCR program was 95°C for 7 min; 35 cycles of 95°C 1 min; 52°C 1 min and 72°C 3 min, followed by an additional 72°C for 10 min; the PCR products were stored at 4°C until use. The length of the PCR product was 1 kb. The extracted DNA samples were also used for real-time PCR analysis.

Western blot analysis

Mouse skin samples were weighed, snap-frozen in liquid nitrogen, and ground in a prechilled tissue grinder. The samples were transferred in 0.5- to 1.0-ml cold lysis buffer into 1-ml glass tissue homogenators (Wheaton Science Products, Millville, NJ, USA) for homogenation. The lysates were sonicated for 10 s at output 20 (Sonifier 250; Branson Ultrasonics Corp., Danbury, CT, USA) and centrifuged at 12,000 g, 4°C for 10 min; the supernatants were collected and protein concentrations were determined by a Bradford Assay with Bio-Rad Protein Assay Reagent. SDS loading buffer (4×) was added to the supernatant and heated at 65°C for 5 min; then protein (30 μg) of each supernatant sample was resolved in a 10% SDS-PAGE. The proteins in the gel were electro-transferred onto nitrocellulose membrane. The membrane was blocked in 5% nonfat Carnation skim milk in TBST (10 mM Tris, pH 8.0, 150 mM NaCl, and 0.05% Tween 20) for 1 h at room temperature, and then immunoblotted with rabbit anti-hCXCR2 primary antibody followed by staining with IRDye conjugated donkey anti-rabbit IgG. The membrane was then scanned in a Li-Cor Odyssey scanner.

RT-PCR

RNA extracted from mouse skin, lung, and liver was treated with DNase I using the Message RNA Clean kit (Gen Hunter Corp., Nashville, TN, USA). The DNA-free RNA was used for RT-PCR. The primers were as follows: K14hCXCR2: 5′ primer: 5′-AATTCTGGCTGGCGTGGAAAT-3′ and 3′ primer: 5′-GGGGCAGGATCCGTAACGCA-3′; β-actin: 5′-primer: 5′-CCACCAGACAACACTGTGTTG-3′, 3′-primer: 5′-AGAGGTATCCTGACCCTGAAG-3′. Typical RT-PCR conditions were as follows: after 7 min of initial denaturation, 35 cycles of denaturing at 95°C for 1 min, annealing at 52°C for 1 min and extension at 72°C for 3 min, followed by 10 min of final extension at 72°C. Ten microliters of each PCR product were analyzed by 1.2% agarose gel electrophoresis with ethidium bromide staining.

Punch biopsy

K14hCXCR2 mutant and wild-type transgenic mice were subjected to punch biopsy under a laminar flow hood. The mice were put to sleep by injection of anesthetic [0.1 mg ketamine (Fort Dodge Animal Health, Fort Dodge, IA, USA) and 0.01 mg xylamine (Ben Venue Laboratories, Bedford, OH, USA)/g mouse weight]. The backside of each mouse was shaved, and the fine hair was further removed with a hair remover. After washing the bare skin with betadine solution followed by 70% alcohol, two or four full-thickness wounds were created at the dorsal area of each mouse by sterile, single-use 4 mm AcuPunch (Acuderm, Inc., Ft. Lauderdale, FL, USA). The wounds were cleaned with Nolvasan S disinfectant (Fort Dodge Animal Health). Liquid bandage (Guard, Beiersdorf, Inc., Wilton, CT, USA) was sprayed on the wounds to prevent the tear of the wounds due to scratching. Buprenex (0.1 to 0.2 μg in sterile saline, Reckitt Benckiser Pharmaceuticals Inc., Richmond, VA, USA) was injected intramuscularly before mice awoke from surgery to reduce the pain. Mice were housed with 4 mice per cage postwounding. On postwounding days 3, 5, 7, and 10, mice were killed by CO₂ inhalation. The wounds were excised and either used for RNA extraction or were fixed, embedded, and sectioned for immunohistochemistry.

Myeloperoxidase assay

Punch biopsy was performed on the mice according to the above procedure. Wound beds surrounded by a very thin margin of nonwounded skin (1 to 1.5 mm) were collected at days 1, 2, and 3 postwounding. Six wounds were collected for each genotype for each time point. Six pieces of skin from unwounded areas were collected from each genotype as a control. Each tissue was ground in liquid nitrogen, then homogenized on ice in 1 ml of homogenization buffer containing 0.5% hexadecyl trimethyl ammonium bromide (HTAB), 0.5 mM EDTA, 0.44 M KH₂PO₄, and 0.062 M K₂HPO₄ (pH=6.0). The homogenized mix was transferred into a 15-ml conical tube, then sonicated on ice for 15 s. The sonicated mix was transferred into an Eppendorf tube, then centrifuged at 14,000 rpm at 4°C for 20 min. The cleared lysates were transferred into a new Eppendorf tube for the myeloperoxidase (MPO) assay. The MPO assay was performed by adding 50 μg protein to 500 μl MPO assay buffer containing 0.0878 M KH₂PO₄, 0.0123 M K₂HPO₄, 0.1668 mg/ml o-dianosidine dihydrochloride (ODH) and 0.0005% H₂O₂. The reaction was initiated with the addition of H₂O₂. The OD₄₉₀ was read every 10 s for 2 min with Beckman-DU 7000 (Beckman Coulter, Fullerton, CA, USA).

Wound resurfacing rate measurement

Punch biopsies were performed on the mice according to the above procedure. Wounds were collected at days 3, 5, 7, and 10 postwounding. Six wounds were collected for each genotype for each time point. Each wound was centrally bisected, embedded, sectioned, and stained with Gomori’s trichrome. To measure the wound-resurfacing rate, the lengths of the new epithelia on both edges of the wound (starting from the innermost edge of the hair follicle to the ending edge of the new epithelia) were measured using Image-Pro Plus 5.1 software (Media Cybernetics Inc., Silver Spring, MD, USA) and added together as total new epithelia length. The length of the whole wound was also measured. The resurfacing rate was obtained by dividing the total new epithelia length by the whole length of the wound [17,18,19].

Quantitation of angiogenesis and macrophage density within the wound bed

Punch biopsies were performed on the mice according to the above procedure. Wounds were collected at days 3, 5, 7, and 10 postwounding. Six wounds were collected for each genotype for each time point. Each wound was centrally bisected, embedded, sectioned, and stained. Capillaries were visualized with CD31 antibody, and macrophages were visualized with F4/80 antibody. Three random measurements were taken from each cutting edge (below the innermost hair follicles in the wounded area), the middle area (the central area in between the two inner most hair follicles in the wounded area), and within each wound bed using a defined magnification. Then the percentage of area occupied by capillaries or macrophages were analyzed by the Image-Pro Plus software [20,21,22].

RNA extraction

The RNA from the collected wounds or skin tissue was extracted with the Qiagen RNeasy Fibrous Tissue Mini Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s protocol. Briefly, the tissue was ground in liquid nitrogen, then homogenized followed by digestion in proteinase K solution at 55°C for 10 min. The lysate was centrifuged for 3 min at 10,000 g, and the supernatant was transferred to a new tube, mixed with 0.5 volume of 100% ethanol and loaded to the RNeasy Mini Spin Column. After washing the column once, the in-column digestion of DNA was performed by adding DNase I, and the column was incubated for 15 min at room temperature. After the column was washed two times with washing buffer, 30-50 μl RNase-free water was added to the column; then the column was centrifuged, and the RNA solution sample was collected.

Real-time PCR

DNA Analysis (Quantitative PCR Genotyping)

Multiplex quantitative PCR (qPCR) was performed for both CXCR2 and β-actin genes simultaneously in a 20-μl reaction containing 50 ng of genomic DNA, 1× iQ SuperMix (Bio-Rad Laboratories, Hercules, CA, USA), 900 nM of each primer, and 100 nM of each 5′ nuclease probe (Table 1). Amplification was carried out using the iCycler, IQ Detection System (Bio-Rad Laboratories) with the following cycling parameters: denaturation (95°C for 10 min) followed by 40 cycles of a two-step PCR (95°C for 15 s, 60°C for 60 s). Samples were analyzed in triplicate. Quantitation was performed through comparisons to standard curves generated with known copy numbers of the defined amplicons. Copy number of each transgenic animal was determined by normalizing the CXCR2 content to β-actin.

TABLE 1.

Primers and Probes Used for the Real-Time PCR

Primers and Probes		5′ to 3′
β-actin	Forward	ACGGCCAGGTCATCACTATTG
	Reverse	CAAGAAGGAAGGCTGGAAAAGA
	Probe	[AminoC6+JOE]-CAACGAGCGGTTCCGATGCCCT-[BHQ-1]
hCXCR2	Forward	CCGCTCCGTCACTGATGTCT
	Reverse	GGCAAGGTCAGGGCAAAGA
	Probe	[6-FAM]-TGCTGAACCTAGCCTTGGCCGACC-[BHQ-1]

RNA Analyses (qRT-PCR)

qRT-PCR analysis (iCycler, iQ Detection System; Bio-Rad) was performed to examine the expression of the human CXCR2 transgene in dermal wounds of experimental animals. RNA (50 ng) was amplified in a 20-μl reaction containing 1× iScript One-Step RT-PCR Supermix (Bio-Rad), 900 nM of each primer, and 100 nM of each 5′ nuclease probe (Table 1). Amplification was carried out using the iCycler, IQ Detection System (Bio-Rad Laboratories) with the following cycling parameters: cDNA synthesis (50°C, 10 min), denaturation (95°C for 5 min) followed by 40 cycles of a two-step PCR (95°C for 10 s, 60°C for 30 s). Samples were analyzed in triplicate. Quantitation was performed by the Comparative Critical Threshold Method (User Bulletin 2, 2002; Applied Biosystems, Foster City CA, USA) to determine the relative changes in transgene expression levels as compared with a pool of RNA isolated from normal skin of 10 animals housing the wild-type CXCR2 transgenic mice. All data were normalized for β-actin expression. Validation of assay efficiency was confirmed by examining amplification profiles of 10-fold serial dilutions of the pooled control for each primer/probe set examined.

Primers and Probes

The following primers and probes were used: β-actin forward primer 5′ ACGGCCAGGTCATCACTATTG 3′, reverse primer 5′ CAAGAAGGAAGGCTGGAAAAGA 3′, probe 5′ [AminoC6+JOE]-CAACGAGCGGTTCCGATGCCCT-[BHQ-1] 3′; hCXCR2 forward primer 5′ CCGCTCCGTCACTGATGTCT 3′, reverse primer 5′ GGCAAGGTCAGGGCAAAGA 3′, and probe 5′ [6-FAM]-TGCTGAACCTAGCCTTGGCCGACC-[BHQ-1] 3′.

Immunohistochemistry

The wounds collected for immunohistochemistry (IHC) were fixed in 4% paraformaldehyde/PBS (pH 6.7) for 24 h; then fixed tissues were processed in the Vanderbilt University School of Medicine Mouse Pathology and IHC Core Lab for Trichrome staining, CD31 staining (goat anti-PECAM-1, 1:400 dilution, Santa Cruz Biotechnology, Santa Cruz, CA, USA) and F4/80 staining (rat anti-mouse F4/80, 1:100 dilution, Serotec Inc., Raleigh, NC, USA). Mouse monoclonal antibodies to E-cadherin and desmoplakin were kind gifts from Dr. James K Wahl III (University of Nebraska Medical Center, College of Dentistry, Lincoln, NE, USA). The hCXCR2, mCXCR2, E-cadherin, and desmoplakin staining was performed according to our standard procedures using antibodies that have been described previously with specificities verified [23,24,25]. The IHC protocol involved deparaffinization/hydration of sections by three xylene washes for 10 min each, two washes of 100% ethanol for 10 min each, and two washes of 95% ethanol for 10 min each. The sections were subsequently washed twice in dH₂O for 5 min each followed by one wash in 1× PBS for 5 min. For antigen unmasking, sections were heated at 100°C in 10-mM sodium citrate buffer (pH 6.0) for 1 min at full power followed by 9 min at medium power, and then the slides were cooled for 20 min. After antigen unmasking, sections were washed in dH₂O three times for 5 min each. Sections were then incubated in 1% H₂O₂ for 10 min and washed in dH₂O three times for 5 min each followed by one PBS wash for 5 min. Sections were blocked for 1 h at room temperature prior to aspiration of the blocking solution, application of the diluted primary antibody, and incubation at 4°C overnight. The primary antibody was removed and sections washed in 1× PBS three times for 5 min each. Diluted secondary antibody was incubated at room temperature for 30 min. Secondary antibody was removed, and sections were washed in 1× PBS three times for 5 min each. ABC reagent was added and incubated at room temperature for 30 min. ABC reagent was removed, and sections were washed in 1× PBS three times for 5 min each. AEC substrate (Vector Laboratories, Burlingame, CA, USA) was added, and sections were incubated at room temperature for 10-30 min, until the color developed. As soon as visible color appeared, the sections were immersed in dH₂O. Sections were counterstained in hematoxylin for 1 min, washed in dH₂O, and then mounted with cover slips.

Culture of keratinocytes from newborn mice

Twenty-four to forty-eight-hour-old newborn pups were killed by CO₂ inhalation, then cleaned with betadine followed by 70% cold ethanol. The skin tissues of chest, abdominal, and dorsal areas were removed and then placed dermal side down in petri dishes. Then 0.25% trypsin solution was added, and the whole skin was digested at 4°C for overnight. The epidermal sheet was separated from underlying dermis, minced with scalpel blade, and further digested by stirring the diced epidermal sheet in 0.25% trypsin solution in a sterile trypsinization flask 30 min at room temperature. The single cell suspension was obtained by filtering the digestion mixture through a nylon cell strainer. The cells were collected by centrifugation, washed with PBS, resuspended in keratinocyte growth medium containing 1% dialyzed FBS, and cultured in 10-cm tissue culture dishes. Twenty-four hours later, the medium and floating cells were removed, and the attached keratinocytes were refed with defined keratinocyte growth medium (EpiLife medium supplemented with 0.06 mM Ca²⁺, HKGS, and antibiotics, EpiLife Medium and HKGS; Cascade Biologics, Inc., Portland, OR, USA).

Migration assays after scratch wounding keratinocyte monolayers

After the cultured keratinocytes from transgenic mice reached 100% confluence, a scratch wound was made in each well, wells were washed twice with Hanks Balanced Salt Solution (HBSS) (Mediatech, Inc., Herndon, VA, USA) and cells were cultured in Epilife growth medium at 37°C in a water jacketed-CO₂ incubator. Images of scratch wounds were obtained at different time intervals and the areas of wound open space were measured by Image Pro software. Comparisons were made between the wound closure rates of mouse keratinocytes expressing wild-type human CXCR2 transgene and that of keratinocytes expressing human CXCR2/331T/LL/AA/IL/AA mutant gene. The experiments were performed a minimum of 3 times.

Phospho-AKT and actin localization in wounded monolayers of keratinocytes

For phospho-AKT staining of keratinocyte cultures, a monolayer of keratinocytes growing on collagen-coated coverslips was wounded by scratching the monolayer with a pipette tip. After incubation at 5% CO₂ and 37°C for certain periods of time, cells on coverslips were fixed in 3.8% paraformaldehyde for 10 min at room temperature. Cells were washed three times with PBS, then permeabilized by incubating with a 0.1% Triton X-100/PBS solution for 5 min at room temperature. After blocking with 1% BSA in PBS for 20 min at room temperature, cells were immunostained with anti-phospho-Akt antibody (Cell Signal Technology, Danvers, MA, USA) followed by a FITC-conjugated goat anti-rabbit antibody, and lastly, cells were stained with rhodamine conjugated phalloidin (Molecular Probes, Eugene, OR, USA) for F-actin staining.

Immunofluorescence staining and confocal microscopy

Cells were grown on glass coverslips, placed in serum-free media, and stimulated with vehicle (0.1% BSA/PBS) or 100 ng/ml MIP-2 diluted in 0.1% BSA/PBS at 37°C for 10 min or 30 min. Cells were fixed in 4% paraformaldehyde for 10 min. Cells were then permeabilized in 0.2% Triton X 100/PBS 5 min and blocked in 10% normal donkey serum 30 min (Jackson Immunoresearch Laboratories, Inc., West Grove, PA, USA). Primary antibodies against human CXCR2 [rabbit polyclonal, generated and affinity purified and characterized for specificity by Richmond Laboratory, as described previously (23)] were added to cells and incubated for 2 h. Cells were washed three times with 0.1% Tween 20/PBS followed by incubation with secondary antibodies against rabbit (Cy 2-conjugated donkey anti-rabbit, Jackson Immunoresearch Laboratories, West Grove, PA, USA) for 1 h. Cells were washed three times with 0.1% Tween 20/PBS. Coverslips were mounted using ProLong Gold anti-fade reagent (Molecular Probes, Invitrogen, Carlsbad, CA, USA). Confocal images were acquired using a LSM-510 Meta laser scanning microscope (Carl Zeiss, Thornwood, NY, USA) with a 63×1.3 numerical aperture oil immersion lens.

RESULTS

Development of hCXCR2 transgenic mice where transgene expression is directed to keratinocytes using the k14 promoter/enhancer

Transgenic mouse lines expressing either wild-type hCXCR2 or mutant hCXCR2 331T LL/AA/IL/AA under the direction of the keratin-14 promoter enhancer were developed as described in Materials and Methods with the assistance of the Transgenic Mouse Core Facility at Vanderbilt University School of Medicine. Two founder lines for K-14 hCXCR2 WT, and three founder lines for K-14 hCXCR2 331T/LL/AA/IL/AA were generated.

The transgene expression levels are different among the different transgenic founders

With real-time PCR, transgene relative copy numbers were obtained from DNA extracted from mouse tails. The mRNA expression levels of transgene were obtained from RNA extracted from intact nonwounded skin samples. For the two founders of each genotype we studied, the relative DNA copy numbers were much lower for K14hCXCR2 WT founder 10 and K14hCXCR2 331T/LL/AA/IL/AA founder 15 as compared with other founders (Fig. 2a). As observed with the transgene relative copy numbers, the RNA expression level of skin was also extremely low on these two founders (Fig. 2b), likely due to the integration of the transgene into a silent site for K14-hCXCR2 WT founder 10. In the case of founder 15 for K14hCXCR2 331T/LL/AA/IL/AA, the level of transgene DNA was extremely low. Because the K14hCXCR2 331T/LL/AA/IL/AA founder 17 exhibited a visible phenotype, an additional K14hCXCR2 331T/LL/AA/IL/AA founder 35 was generated, which also exhibited high expression of hCXCR2 and a visible phenotype. With real-time PCR, the DNA copy number of founder 35 is 0.92 times as high as founder 17, and the mRNA expression level is 1.33 times higher than for founder 17. We focused our analysis on the founders where transgene expression was significant: K14hCXCR2 WT founder 8 and K14hCXCR2 331T/LL/AA/IL/AA founder 17 and 35. As we expected, there is no detectable copy number and RNA expression from nontransgenic mice.

Fig. 2.

The transgene expression levels are different among the different transgenic founders. With real-time PCR, transgene relative copy numbers were obtained from DNA extracted from mouse tails; transgene expression levels were obtained from RNA extracted from intact skin samples or wound samples. (a) Transgene DNA level relative to actin DNA for the transgenic founders. (b) Skin transgene mRNA expression levels for the transgenic founders.

Verification of transgene expression

The expression of hCXCR2 transgenes in the skin were confirmed by RT-PCR (Fig. 3a) and Western blot (Fig. 3b). The RNA samples extracted from the skin, lung, and liver were analyzed by RT-PCR, and only RNA from the skin expressed the hCXCR2 transgene (Fig. 2a). Skin proteins extracted from transgenic and nontransgenic littermates were subjected to Western blot and immunodetection with antibody to hCXCR2. Human CXCR2 proteins were detected only in the skin from transgenic mice (Fig. 3b). These data indicate that the hCXCR2 is targeted appropriately to the skin.

Fig. 3.

Detection of hCXCR2 expression in the transgenic mouse skin. (a) RT-PCR was performed on RNA of skin, lung, and liver extracted from K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice and nontransgenic littermates. The upper gel picture shows hCXCR2 mRNA and the lower gel picture shows expression of mRNA with β-actin primers. Lanes 1, 3, 5 represent RNA of skin, lung, and liver from transgenic mice, respectively; Lanes 2, 4, 6 represent RNA of skin, lung, and liver from nontransgenic mice respectively. (b) Western blot analysis was performed to detect hCXCR2 expression in the skin samples from K14-hCXCR2 WT and K14-hCXCR2 331T/LL/AA/IL/AA transgenic mice. β-actin expression was shown as control. (c) Immunohistochemical detection of hCXCR2 and mCXCR2 protein expression postwounding in transgenic mice. Excised wounds were embedded in paraffin, then prepared for immunohistochemical analysis using antibodies for hCXCR2 and mCXCR2 as described in Materials and Methods. Expression of both hCXCR2 and mCXCR2 was observed, but expression of hCXCR2 was dominant in the transgenic mouse wounds. Photomicrographs shown are with ×10 and ×40 magnifications. Rabbit IgG and goat IgG isotopic antibody staining was used as nonspecific controls for immunoreactions.

The detection of hCXCR2 gene expression of the skin and wounds of the transgenic mice by immunohistochemistry staining

Immunohistochemical staining was performed on the tissue sections with rabbit anti-hCXCR2 antibody (rabbit polyclonal, generated from our laboratory as previously characterized [24, 27] and affinity purified) to examine the expression of hCXCR2 in the skin and wounds at different time points after wounding. We also examined the level of expression of the murine CXCR2 by immunostaining using an goat anti-mCXCR2 (provided by Dr. Robert Strieter, UCLA School of Medicine and previously characterized by Addison et al.) [23]. Expression of both hCXCR2 and mCXCR2 were observed, but the expression of hCXCR2 was dominant in the wounds. Figure 3c shows the staining on the wounds from the transgenic mice and nontransgenic control littermates. Human CXCR2 was expressed strongly in the epidermis undergoing resurfacing and mouse CXCR2 was prominent in the vascular bed underlying the resurfaced epidermis. We observed equivalent levels of hCXCR2 immunostaining in the epidermal keratinocytes of K14-hCXCR2 WT and K14-hCXCR2-331T LL/AA/IL/AA transgenic mice, indicating that even through the level of mRNA expression in the K14-hCXCR2-331T/LL/AA/IL/AA mice was elevated compared with K14-hCXCR2 WT transgenic mice, the levels of CXCR2 protein expression were not elevated.

Visual phenotype of transgenic mice

The K14hCXCR2 WT transgenic mice were visually not different from nontransgenic control littermates. However, progeny from the K14hCXCR2 331T/LL/AA/IL/AA transgenic founder 17 and 35 exhibited a visible phenotype. Interestingly, the K14hCXCR2 331T/LL/AA/IL/AA founder mice were born with tails of normal length, but for 99% of the transgenic mice, three to eight days after birth their tails fell off, leaving only a short tail stub (Fig. 4, a–c). From day 3, tissue degeneration started between caudal somites and was accompanied by degeneration of some of the bone and connective tissue. Gradually, most of the bone and cartilage distal to the constriction were replaced with stromal tissue heavily infiltrated with inflammatory cells. In addition, the cohesivity of the vascular structures distal to the constriction was progressively lost. Analysis of the lesion site in the tail revealed coagulation in enlarged vessels and marked edema that eventually led to the loss of distal tail (Fig. 4d). Moreover, 66% (29 out of 44 transgenic mice observed had this skin problem) of these mice exhibited varying degrees of skin problems on the face, neck, ears, back, and legs. In the affected areas of the skin, there was hair loss, fluids from the affected area were seeping out and often bloody, and as the areas healed over, there was visible scaring. The skin problem occurs as early as 2 mo of age. Histological analysis of the spontaneously arising skin lesions of the K14hCXCR2 331T/LL/AA/IL/AA transgenic mice revealed an increase in the number of sebaceous glands, enlargement and thickening of the hair follicles, enlarged, abundant blood vessels, and a thickening of the epidermis. Trichrome staining of these lesions to reveal connective tissue showed a marked reduction in the connective tissue in the skin lesions of the K14hCXCR2 331T/LL/AA/IL/AA mutants (Fig. 5, d and e).

Fig. 4.

(a) A litter of 2-day-old pups from K14hCXCR2 331T/LL/AA/IL/AA founder 17. Both transgenic and nontransgenic newborn mice have normal length of tails and appear similar at 2 days postbirth. (b) 8 days postbirth, the tails of the two K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice have degenerated, while the two nontransgenic mice have normal tails. (c) Adult K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mouse (left) and nontransgenic litter mate (right). (d) H&E staining of the tail tissue from K14hCXCR2 331T/LL/AA/IL/AA founder 17 mice harvested at post-natal days 5 (A and B), 7 (C and D) and 8 (E and F) during the time of tail degeneration and detachment. Magnifications are ×5 and ×80. (e) Histology of tail sections from K14hCXCR2 WT and K14hCXCR2 331T/LL/AA/IL/AA mice. Immunohistochemistry was performed with antibodies specific to E-cadherin or desmoplakin in tail sections from K14hCXCR2 WT mice and K14hCXCR2 331T/LL/AA/IL/AA mice. (A and B) Uniform localization of E-cadherin is demonstrated within all layers of the normal epidermis, as well as in hair follicles in K14hCXCR2 WT mice. (C and D) Down-regulation of E-cadherin was noticeable within all layers of the normal epidermis, as well as in hair follicles in K14hCXCR2 331T/LL/AA/IL/AA mice. (E and F) Uniform localization of desmoplakin is demonstrated within all layers of the normal epidermis, as well as in hair follicles in K14hCXCR2 WT mice. (G and H) Down-regulation of desmoplakin was detected within all layers of the normal epidermis, as well as in hair follicles in K14hCXCR2 331T/LL/AA/IL/AA mice. Magnifications are ×20 and ×40.

Fig. 5.

Cutaneous lesions on the K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice. Approximately 50% of these mice exhibited varying degrees of skin problems on the face, neck, ears, back, and legs. In the affected areas of the skin, there was hair loss; fluid leakage from the affected area was often bloody, and as the areas healed, there was visible scarring. The skin problem occurs as early as 2 mo of age. (a and b) Pictures of the cutaneous lesions on the K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice. (c and d) H&E staining on the normal skin and the skin with cutaneous lesions of the K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice. (e and f) Trichrome green staining on the normal skin and the skin with cutaneous lesions of the K14hCXCR2 331T/LL/AA/IL/AA founder 17 transgenic mice. Magnification ×20.

Detection of E-cadherin and desmoplakin in tail sections from K14hCXCR2 WT and K14hCXCR2 331T/LL/AA/IL/AA mice

The K14hCXCR2 331T/LL/AA/IL/AA mice presented with detached tails, skin erosions on face, neck, ear, and back, as well as thickening of the epidermis. Therefore, we asked whether there was a correlation between the detached tails and skin lesions of transgenic mice and a loss of cell adhesion markers such as the adherens junction marker E-cadherin and the desmosomal marker desmoplakin. We examined localization of E-cadherin and desmoplakin by immunohistochemistry in tail sections from K14hCXCR2 WT and K14hCXCR2 331T/LL/AA/IL/AA mice. Immunohistochemical staining was performed on tissue sections with mouse anti-E-cadherin and mouse anti-desmoplakin. Immunohistochemical analysis of markers associated with keratinocyte intercellular junctions revealed that E-cadherin immunoreactivity was detected throughout all layers of the epidermis, as well as in hair follicles of K14hCXCR2 WT mice (Fig. 4e, A and B). In contrast to uniform expression of E-cadherin in wild-type mice, E-cadherin immunoreactivity was absent and therefore down-regulated in K14hCXCR2 331T/LL/AA/IL/AA mice (Fig. 4e, C and D). Similarly, the expression of desmoplakin was uniform throughout all of the layers of the epidermis, as well as in hair follicles of K14hCXCR2 WT mice (Fig. 4e, E and F), as compared with a down-regulation of desmoplakin in K14hCXCR2 331T/LL/AA/IL/AA mice (Fig. 4e, G and H). These results indicate that the loss of E-cadherin in the adherens junction and the loss of desmoplakin in the desmosome in tail sections from K14hCXCR2 331T/LL/AA/IL/AA mice occur in conjunction with the loss of cell-cell adhesion in relation to the detached tails and skin lesion phenotype in the transgenic mice.

Wound healing studies

Since the phenotype of the transgenic mice expressing mutant hCXCR2 suggested that there were defects in the maintenance of the epidermis of the skin and tail, we proceeded to perform excision cutaneous wound healing experiments. Moreover, prior work from our group has demonstrated that loss of CXCR2 resulted in delayed wound healing responses to excision wounds and chemical burn wounds. Excision biopsy (punch biopsy) is a widely used technique for cutaneous wound healing studies. Punch biopsies were performed on nontransgenic and transgenic mice expressing wild-type CXCR2, mutant CXCR2 (331T/IL/AA/LL/AA) mice where expression of the transgene is directed by the Keratin-14 promoter/enhancer. After punch biopsy, the wounds were collected at 3, 5, 7, and 10 days postwounding. The wounds were either snap frozen or fixed in paraformaldehyde and embedded for further experiments to characterize neutrophil infiltration, wound closure rate, macrophage infiltration, and capillary density. After the wounds were collected, the fixed specimens were examined after trichrome staining to quantify the percentage of epithelial resurfacing. Neutrophil infiltration was assessed based upon MPO activity. Macrophage density was determined by microscopic analysis and quantitation of fixed, sectioned wound bed tissues stained with the F4/80 antibody that recognizes macrophages. Capillary density measurements were based upon quantitation of CD31 (PECAM-1) immunostaining of endothelial cells from fixed sectioned wound bed tissues.

Neutrophil infiltration into wound bed

To access neutrophil infiltration into the wound bed over time, the MPO assay was performed on snap-frozen wounds. MPO is an enzyme located in the azurophilic granules of neutrophils that can produce very toxic compounds that aid in the killing of bacteria, fungi, and tumor cells [26]. MPO is recoverable from skin in soluble form, and MPO activity is directly related to the extent of neutrophil infiltration and therefore can be used as a surrogate marker for neutrophil infiltration. In our study, the MPO activities of nontransgenic and the two genotypes of transgenic mice were determined. Four wounds were made on each mouse by a 4-mm-diameter AcuPunch (Acuderm Inc., FL, USA); then the wounds were collected on postwounding days 1, 2, and 3. Four pieces of skin were collected from unwounded mice used as control for the MPO activity assay. Samples collected from three mice of each genotype on post wound day 1, 2, and 3 were used for the MPO assay. For the unwounded skin samples, the MPO activities were low for all genotypes. For the samples collected from postwound days 1 and 2, the MPO activities were elevated, but no significant difference was observed between nontransgenic, K14hCXCR2 331T/LL/AA/IL/AA mutant and K14hCXCR2 WT, though the nontransgenic mice exhibited a higher level of MPO activity at postwound day 3 than did the transgenic mice (Fig. 6). Thus, during the first 3 days after wounding, the expression of mutant hCXCR2 331T/ LL/AA/IL/AA in keratinocytes did not affect neutrophil infiltration into the wound bed as compared with hCXCR2 WT.

Fig. 6.

MPO activity of unwounded skin and cutaneous excision wound samples. MPO activity from unwounded skin was low for all genotypes; for the cutaneous excision wound samples collected from postwound days 1, 2, and 3, the MPO activities were elevated, but no significant difference was observed between K14hCXCR2 331T/LL/AA /IL/AA mutant and K14hCXCR2 wild-type (for postwound day 3; *, P=0.0574, Student’s t test).

CXCR2 mutation does not have a major influence wound closure rate in vivo

When the resurfacing rate of excision wounds were compared between nontransgenic mice, K14hCXCR2 WT transgenic mice, and K14hCXCR2 331T/ LL/AA/IL/AA transgenic mice, significant differences were not observed in the wound closure rate with one exception. Postwound day 5 transgenic mice expressing mutant hCXCR2 closed slightly faster than wounds on hCXCR2WT transgenic mice (Fig. 7). For these studies, 24 wounds from each genotype were examined in two independent experiments.

Fig. 7.

hCXCR2 status does not influence wound closure rate. Excision wounds were made in nontransgenic, K14hCXCR2 WT transgenic, and K14hCXCR2 331T/LL/AA/IL/AA transgenic mice. After the wounds were collected, the fixed specimens were stained with trichrome and examined microscopically to quantify the percentage of epithelial resurfacing as described in Materials and Methods. No significant differences were observed in the wound closure rate between nontransgenic, K14hCXCR2 WT transgenic, and K14hCXCR2 331T/LL/AA/IL/AA transgenic mice.

CXCR2 status influences the timing of peak capillary density in the wound

To determine whether there were differences in capillary density in wounds from the various transgenic mice, sections of wounds at postwound day 3, 5, 7, and 10 were stained with CD31 antibody, which detects PECAM in endothelial cells. When counting the capillary density, three areas at each edge of the wound and in the middle of the wound were selected for counting at a magnification of ×40. Figure 8a shows that the capillary density for the K14hCXCR2 331T/LL/AA/IL/AA transgenic mice peaked at postwound day 3 and then gradually declined. In contrast, the nontransgenic and K14hCXCR2 WT transgenic mice showed a peak capillary density at day 7 and declined by postwound day 10. The K14hCXCR2 331T/LL/AA/IL/AA founder 17 had a significantly lower capillary density at postwound day 7 compared with the nontransgenic and K14hCXCR2 WT transgenic founder 8 (P<0.05, Student’s t test). The K14hCXCR2 WT transgenic mice exhibited significantly reduced capillary density relative to nontransgenic and to K14hCXCR2 331T/LL/AA/IL/AA transgenic mice on postwound day 10 (P<0.05, Student’s t test). For these studies, 24 wounds from each genotype were examined in two independent experiments. Altogether, the main difference in capillary density observed among the mice of the three genotypes is in the timing of the peak response, as opposed to the magnitude of the response.

Fig. 8.

CXCR2 status influences capillary density and macrophage density in the wound area at different time points after wounding. (a) After the wounds were collected, fixed, sectioned, the endothelial cells were stained with antibody to CD31, and capillary density was quantitated as described in Methods. (b) Macrophages were identified by immunohistochemistry in fixed, sectioned wounds with F4/80 antibody and staining was quantitated as described in Methods. Outer borders of wounds and wound beds were evaluated separately, then combined with inner wounds to access overall density of CD31 or F4/80 staining.

CXCR2 Status Influences Macrophage Density in the Wound Area at Different Time Points

F4/80 immunostaining was analyzed in paraffin-embedded sections of wounds at postwound days 3, 5, 7, and 10 to detect macrophages. When counting the macrophage density, three areas at each edge of the wound and in the middle of the wound were selected for counting at a magnification of ×40. The total sum of the areas was used to obtain the macrophage density for the whole wound. Figure 8b shows that the macrophage density of the wound was considerably reduced in the K14hCXCR2 331T/LL/AA/IL/AA founder 17 on postwound days 5 and 7 compared with K14hCXCR2 WT controls but was not significantly different from the nontransgenic mice on those days. However by postwound day 10, this transgenic founder line expressing mutant hCXCR2 showed increased macrophage density relative to nontransgenic and WT control transgenic mice (P<0.05, Student’s t test). The macrophage density in wounds of transgenic mice expressing mutant CXCR2 peaked at postwound day 3, fell by postwound day 7 then rose again by postwound day 10. This was in contrast to a peak macrophage density in the nontransgenic mice on postwound day 5 and on post wound day 7 for the K14hCXCR2 WT transgenic mice. For these studies, we examined 24 wounds from each genotype in two independent experiments. The reason for the higher macrophage number in the K14hCXCR2WT mice on postwound days 5 and 7 is not apparent but may relate to enhanced production of factors stimulating macrophage recruitment in response to activation of the higher levels of functional CXCR2 on keratinocytes.

Truncation of the carboxyl-terminal domain or mutation of the LL/IL motif of CXCR2 does not impair ligand-mediated receptor internalization

To determine whether the mutant form of CXCR2 displayed properties similar to those observed for this receptor in HEK 293 or HL60 cells, keratinocytes were cultured on glass cover slips, stimulated or not stimulated with ligand for CXCR2 (MIP-2), and the internalization of receptor was examined by confocal immunofluorescence analysis. CXCR2 was detected with rabbit anti-human polyclonal antibody to hCXCR2. Photomicrographs of confocal images were examined. Figure 9 shows the confocal images of immunofluorescence staining of hCXCR2 following Mip-2 stimulation in keratinocytes isolated from newborns of different transgenic mice. Cells were stimulated with vehicle (Untreated) or 100 ng/ml of MIP-2 (the kind gift of Elias Lolis, Yale University School of Medicine) for 10 min and 30 min. Unexpectedly, like hCXCR2 WT, the hCXCR2 331T/LL/AA/IL/AA mutant receptor also underwent extensive internalization after ligand stimulation in the cultured keratinocytes. These data point to any defects in biological response for mice expressing the K14hCXCR2 331T/LL/AA/IL/AA being independent of effects on receptor internalization after ligand stimulation.

Fig. 9.

Confocal images of immunofluorescence staining of hCXCR2 following MIP-2 stimulation of cultured keratinocytes isolated from transgenic mice. Cells were stimulated with vehicle (untreated) or 100 ng/ml of MIP-2 for 10 min and 30 min. Cells were fixed, immunostained for hCXCR2, examined by confocal microscopy, and photomicrographed. Scale bars = 10 μm.

Keratinocyte expression of hCXCR2 331T/LL/AA/IL/AA mutant enhances the rate of wound closure for cultured keratinocytes in a scratch wound assay

Keratinocyte cultures were established from K-14-hCXCR2WT and K-14-hCXCR2 331T LL/AA/IL/AA transgenic newborn pups. Cells were cultured until a confluent monolayer was reached, then scratch wounds were made as described in Materials and Methods. We evaluated the closure of the wounds over a time course of 48 h and took photomicrographs at 0, 24, and 48 h postwounding. Surprisingly, we observed that the wounds from the keratinocytes established from the K-14-hCXCR2 331T/LL/AA/IL/AA mice exhibited slightly enhanced wound closure as compared with keratinocytes established from transgenic mice expressing WT hCXCR2. Because no exogenous MIP-2 was applied during the wound-healing process, the effects observed were likely dependent upon MIP-2 or KC (murine orthologs of CXCL1 and CXCL8) secreted into the culture medium upon wounding (Fig. 10, A and B).

Fig. 10.

(a) Scratch wounds were made in cultured keratinocyte monolayers established from newborn skin from K14-hCXCR2 WT and K14-hCXCR2 331T LL/AA/IL/AA transgenic mice, and the images of the open wound area were taken at 0, 24, and 48 h after wounding (A). The area of open wound space was measured by Image Pro software. The wound area changes relative to original wound at 0 h are shown (B). The graph represents the mean ± sd of more than 3 experiments. (b) Confocal images of immunofluorescence staining of hCXCR2 following scratch wounding of cultured keratinocytes isolated from transgenic mice. Cells were fixed, immunostained for pAKT and phalloidin, and examined by confocal microscopy; then photomicrographs were taken.

Immunostaining of these wounded keratinocytes cultured from K14-hCXCR2 WT and K14-hCXCR2 331T/LL/AA/IL/AA mice during the process of wound closure with phospho-Akt antibody for the localization of activated Akt and phalloidin for F-actin revealed no difference between these two genotypes (Fig. 10b). Interesting, most of the phospho-Akt is concentrated in perinuclear region of keratinocytes and in some cells, even in nuclei. These data suggest that at this time point in the wound repair process, the activated AKT is likely involved in proliferation and gene expression of the wounded keratinocyte monolayer.

DISCUSSION

CXCR2 has been established as a major modulator of the wound-healing response. The CXCR2 ligands (CXCL1, CXCL5, CXCL8) are present in keratinocytes and neutrophils and can be released upon wounding to stimulate the wound-healing response [4, 25, 28]. Targeted deletion of CXCR2 results in delay of the wound-healing response, including delayed epithelial resurfacing, neutrophil migration, and wound bed angiogenesis [29]. Topically applied CXCL8 to human skin grafted onto chimeric mice resulted in enhanced re-epithelialization due in part to enhanced proliferation of keratinocytes [30]. In some biological models, mutated chemokine receptors are hyperactive, resulting in enhanced biological responsiveness, and often ligand independence. This has been demonstrated for CXCR2 mutated in the DRY box to VRY [31]. Moreover, mice that have a targeted deletion of β-arrestin exhibit an enhanced wound-healing response and enhanced neutrophil recruitment into an air pouch filled with CXCR2 ligand [18]

In this report, we have characterized the effect of keratinocyte expression of a mutant form of CXCR2 previously shown to be defective in binding to adaptor molecules AP-2 and β-arrestin and in desensitization [15, 32, 33]. Although we clearly expected to observe differences in the wound-healing rate in transgenic mice expressing WT vs. mutant hCXCR2, this was not our finding. Postexcision wound-resurfacing data and neutrophil migration into the wound bed in the K14hCXCR2 331T/LL/AA/IL/AA transgenic mice were not significantly different from WT control transgenic mice, except for postwound day 5 where mutant CXCR2 transgenic wounds closed faster than WT CXCR2 transgenic wounds. This faster wound closure was supported by in vitro studies using cultured keratinocytes from hCXCR2 331T/LL/AA/IL/AA transgenic mice, in which we observed small but statistically significant enhancement in wound closure as compared with that observed in keratinocyte wounds from hCXCR2 WT transgenic mice. However, localization of phosphospho-AKT in a wounded keratinocyte monolayer was not significantly different in keratinocytes from the transgenic mice expressing mutant vs. WT hCXCR2. Surprisingly, in the in vivo excision wounding experiments, there were some differences in macrophage infiltration and capillary density in the wound bed, even though the mutant receptors were only expressed in keratinocytes and not macrophages or endothelial cells. These data suggest that expression of desensitization and adaptor binding-deficient mutants CXCR2 in keratinocytes may change the chemotactic milieu of the wound bed to influence the timing of cell migration into the wound bed.

For founder lines 17 and 35 of the K14hCXCR2 331T/LL/AA/IL/AA transgenic mice, we observed that in 99% of the transgenic mice, 3 to 8 days after birth, the tails detached, leaving only very short tail “stubs.” Sixty-six percent of the mice expressing this transgene developed different degrees of skin problems on the face, neck, ears, back, and legs. The skin problem occurs as early as 2 mo of age. Histological analysis of the spontaneously arising skin lesions of the K14hCXCR2 331T/LL/AA/IL/AA transgenic mice revealed an increase in the number of sebaceous glands, enlargement of the hair follicles, enlarged, abundant blood vessels, a thickening of the epidermis, and a reduction in the extracellular matrix (Fig. 5, d and e). The tails appeared to undergo restriction between caudal somites followed by degeneration and detachment of the distal portion of the tail. It is interesting that the changes in wound healing response after excision wounding in these transgenic mice expressing mutant CXCR2 were not striking, even though there are spontaneous skin lesions and tail abnormalities in these mice. Thus, it is likely that cytokine profiles induced in response to excision wounding are sufficient to induce wound repair with only minor changes in the timing of the wound-healing processes. Alternatively, the wound-healing responses in the K14-hCXCR2331T/LL/AA/IL/AA transgenic mice may be masked by the coexpression of the functional murine CXCR2 on these keratinocytes. This hypothesis is supported by the observation that unexpectedly, the CXCR2331T/LL/AA/IL/AA receptor expressed in keratinocytes underwent ligand-induced internalization, in great contrast to our prior studies of this mutant receptor expressed in HEK293 cells and differentiated HL-60 cells (14, 32). These unexpected findings could be explained, in part, if heterodimers formed between the mutant receptor and the fully functional murine CXCR2, such that the receptor dimer continues to internalize and signal through partnering with murine CXCR2. In addition, the level of expression of the mutant receptor might perhaps countereffect expression of an altered function receptor in these excision wound-healing studies. In the future, experiments using fluorescence resolution eEnergy transfer (FRET) analysis to examine the dimerization of murine and human CXCR2 in these transgenic mice will be the subject of another manuscript. It will also be useful to express the K14-hCXCR2 331T LL/AA/IL/AA transgene on the CXCR2−/− background and compare the phenotype of these mice with that of the K14-hCXCR2 331T/LL/AA/IL/AA on the mCXCR2 +/+ background. These experiments will add additional clarity to our understanding of how expression of this mutation of CXCR2 in keratinocytes affects the homeostasis of skin and tail in addition to effects on excision wound healing.

Though transgene mRNA expression was elevated in mice expressing the hCXCR2 331T/LL/AA/IL/AA compared with WT hCXCR2 transgenic mice, the level of hCXCR2 protein expression (mutant or wild-type) is not elevated based upon Western blot and immunohistochemical staining. This disconnect between mRNA expression and expression of mutant proteins is not uncommon. Therefore, it is unlikely that the observable phenotype is related to the overexpression of the mutant form of hCXCR2 but rather is associated with expression of the mutant CXCR2 transgene. These data suggest that the LL/IL motif and carboxyl terminal domain of CXCR2 are important for CXCR2 tissue maintenance. Moreover, similar experiments performed in two founder lines of K14-transgenic mice expressing a desensitization defective [34] truncated CXCR2 (CXCR2 331T) failed to show the tail and skin phenotype exhibited by the K-14 hCXCR2 331T LL/AA/IL/AA transgenic mice (data not shown). These data suggest that mutation of the LL/IL motif in the carboxyl-terminal domain of hCXCR2, is responsible for the tail and skin phenotype, rather than changes in desensitization of the receptor brought on through truncation of the carboxyl-terminal domain that encodes the serine residues phosphorylated in response to ligand. The timing and location of the lesions that arise in conjunction with the expression of the mutant receptor suggest that there must be a second factor temporally working in concert with the mutation of CXCR2 responsible for the phenotype. One possibility is that at times in the normal maintenance of the epidermis, when expression of the murine wild-type receptor declines, then the abnormal phenotype develops due to reliance of chemokine function based solely on the expression of the mutant form of CXCR2

A number of genetic mutations previously were reported to be associated with reduction in tail length or skin lesions. Homozygous mutations in the Brachyury gene, also referred to as T for tail, can have profound effects on the development of the notochord. Mice heterozygous for mutation in this gene exhibit blunt tails [35, 36]. This T gene encodes a transcription factor [37], which is modulated by a number of Brachyury-modifier genes (Brm1 and Brm2) or Brachyury-interacting genes [38]. The abnormal feet and tail (Aft) gene is located in the Brm1 region. Aft mutant mice display kinky tails, syndactyly in the hind-limb, and late onset hair loss. Interestingly, mice heterozygous for both Aft and T mutation are often tailless [39]. Histological analysis of the skin lesions in Aft mice revealed reduction in hair follicles and increased mast cells [39].

Unlike the Aft mice, we saw no evidence of syndactyly in K14hCXCR2 331T/LL/AA/IL/AA transgenic mice, though the tail loss was common to both Aft/T mutation and K14hCXCR2-331T LL/AA/IL/AA mice. While chemokine receptors do not directly modulate transcription, they do induce signal transduction cascades that activate transcription. Therefore, it may be possible that expression of a mutant form of hCXCR2 with altered function results in altered gene expression leading to tail shortening and skin lesions. The tail shortening process in hCXCR2 mutant transgenic mice began with tail restriction, was accompanied by remodeling of the cartilage tissue, replacement with stromal tissue with inflammatory cells, and enlargement of vessels with edema. The skin lesions occurring in these transgenic mice also were marked by dilation of the blood vessels and inflammatory cell infiltration and thickening of the epidermis.

On the basis of the skin and tail lesion phenotype of K14hCXCR2 331T/LL/AA/IL/AA mice, we postulated that there might be some defect in barrier function of the epidermis. E-cadherin is normally expressed throughout the epidermis and is the main cadherin responsible for cell-cell adhesion between keratinocytes [40]. The loss of E-cadherin is embryonic lethal; however, information from E-cadherin conditional knockout mice demonstrated that E-cadherin plays a role in keratinocyte adhesion, epidermal thickness, and differentiation [41, 42]. In addition, desmoplakin knockout mice are also embryonic lethal; however, desmoplakin-conditional epidermal knockout resulted in mice with peeling skin, as desmosomes in these mice were found to lack attachment to the keratin intermediate filaments [43, 45]. In relation to the desmosome complex, other mice that exhibit skin lesions and or tail generation similar to that we observed are the Desmocollin 1(Dsc 1) knockout mice [44], as well as transgenic mice expressing mutant desmoglein 3 (Dsg 3) under the transcriptional control of the K14 promoter/enhancer [45]. The Dsc1 knockout mice have flaky skin and a barrier defect in the epidermis, hyperproliferative epidermis with abnormal differentiation, ulcerating lesions with dermatitis, hair loss, and dermal cysts. Mice null for Dsg 3 have skin erosions, oral mucosal lesions, eye lesions, erosion of the snout, balding, and separation of the desmosomes are visible by electron microscopy [46]. In contrast, expression of an amino-terminal deletion mutant of Dsg 3 under the transcriptional direction of the K14 promoter/enhancer resulted in mice with tail degeneration, swelling of paws and digits, and flaky skin. Histological examination of the skin revealed a thickening of the epidermis and widening of the keratinocyte intercellular space due to a loss of desmosomes. This information led us to examine closely the epidermis of K14-hCXCR2 331T LL/AA/IL/AA transgenic mice and wild-type K14 CXCR2 transgenic mice. We asked whether there was a correlation between the detached tails and skin lesions of transgenic mice and a loss of cell adhesion markers such as E-cadherin found in the adherens junction and desmoplakin found in the desmosome. These studies demonstrated a loss of E-cadherin and desmoplakin in K14hCXCR2 331T/LL/AA/IL/AA mice as compared with wild-type mice in the epidermis and hair follicles of tail sections, suggesting that, indeed, expression of the mutant receptor is linked to disturbance in barrier function in some, but not all, areas of the epidermis.

These findings of altered expression of E-cadherin in cells expressing carboxyl-terminal domain truncated chemokine receptor are somewhat similar to our earlier studies, showing that expression of CXCR4 truncated in the carboxyl terminal domain in MCF-7 breast cancer cells resulted in down-regulation of expression of E-cadherin and ZO-1. In addition, microarray studies showed many other changes in gene expression in these cells as compared with MCF-7 cells, expressing equivalent levels of wild-type CXCR4. MCF-7 cells expressing the mutant receptor, but not the wild-type CXCR4 underwent morphological changes like epithelial to mesenchymal transformation and exhibited enhanced motility and enhanced proliferation [47].

In future experiments that will be the subject of another manuscript, it will be interesting to see how mutations of CXCR2 alter the gene expression profile of keratinocytes and stromal cells to produce these profound phenotypic changes noted in our transgenic mice. Since the abnormalities are restricted to specific regions of tail or skin, one would expect that other factors must impact the mutant receptor to result in local distribution of the abnormalities.